Identifying ADHD Individuals From Resting-State Functional Connectivity Using Subspace Clustering and Binary...

Area of Science:

• Neuroscience
• Medical Imaging
• Machine Learning

Background:

• Attention Deficit Hyperactivity Disorder (ADHD) is a prevalent neurodevelopmental disorder.
• Accurate classification of ADHD is crucial for timely intervention and treatment.
• Functional connectivity (FC) analysis offers a promising avenue for understanding brain alterations in ADHD.

Purpose of the Study:

• To develop and validate a novel method for ADHD classification using FC data.
• To enhance the accuracy and reliability of ADHD identification through advanced machine learning techniques.
• To identify specific brain circuits that are discriminative for ADHD.

Main Methods:

• A novel ADHD classification method combining subspace clustering and binary hypothesis testing was developed.
• Partial information from test data was utilized during the training phase for improved generalization.
• A multi-affinity subspace clustering approach was employed for feature extraction and dimensionality reduction.

Main Results:

• The proposed method achieved an average identification accuracy exceeding 90%, outperforming existing state-of-the-art techniques.
• Discriminative functional connectivity (FC) contribution analysis confirmed the reliability and robustness of the classification model.
• The study successfully identified specific brain circuits implicated in ADHD.

Conclusions:

• The developed method demonstrates significant potential for accurate and reliable ADHD classification.
• The findings highlight the utility of FC analysis and subspace clustering in neuroimaging-based diagnostics.
• This research provides valuable insights into the neural underpinnings of ADHD and aids in identifying potential biomarkers.